/obsidian-world

description: Load your identity, values, priorities, and current state in one shot - with progressive context levels to avoid burning tokens
category: vault
triggers_en: ["load context", "what is going on", "where am I", "load my world"]
triggers_es: ["carga el contexto", "qué está pasando", "dónde estoy", "ponme al tanto de todo", "dame el resumen de mi vida"]
triggers_pt: ["carregue contexto", "o que está acontecendo", "onde estou", "carregue meu mundo"]
triggers_zh: ["加载我的完整背景", "告诉我现在的整体状况", "读取我的身份和当前重点", "先了解一下我的情况"]

Use the obsidian-second-brain skill. Execute `/obsidian-world`:

1. Read `_CLAUDE.md` first if it exists in the vault root

2. Load context progressively - start light, go deeper only as needed:

**L0 - Identity (~170 tokens)** Read these files if they exist (search for them if paths differ):

• `SOUL.md` or `About Me.md` - who the user is, communication style, thinking preferences
• `CRITICAL_FACTS.md` - ~120 tokens of always-needed context: timezone, manager, location, company, role
• `CORE_VALUES.md` or `Values.md` - decision-making principles and non-negotiables

**L1 - Navigation (~1-2K tokens)**

• Read `index.md` - the catalog of all vault pages. This tells Claude what exists without loading everything.
• Read recent operation log (last 10 entries): if `Logs/` exists, read today's and yesterday's `Logs/YYYY-MM-DD.md`; otherwise read `log.md`

**L2 - Current State (~2-5K tokens)**

• Read `Home.md` or `Dashboard.md` - current top-level priorities
• Read today's daily note (if it exists) for what's already in progress
• Read the last 3 daily notes for recent momentum and open threads
• Scan active kanban boards for in-progress and overdue items
• Check for session digests from previous conversations (look for "End of Day" or "Session Digest" sections)

**L3 - Deep Context (on demand, ~5-20K tokens)**

• Only load if needed for a specific question or task
• Read active project notes (status: active) for current goals and blockers
• Read full source/reference notes (the concept or knowledge folder per `references/folder-map.md` - wiki-style `wiki/concepts/` or `raw/`, Obsidian-style `Knowledge/`) if the user asks about a specific topic
• Identify key people interacted with recently (last 7 days of daily notes)

3. Present a brief status after L0-L2 (do NOT load L3 unless needed):

• **Who I am to you**: confirm the persona and communication style
• **Your current priorities**: top 3-5 active threads (from index.md + boards)
• **Open threads from last session**: anything unfinished (from operation log + daily notes)
• **Overdue / needs attention**: tasks or projects that are stale
• **Today so far**: what's already logged today

Keep output concise - this is a boot-up sequence, not a report. The user should glance at it and say "yes, Claude is up to speed" and start working immediately.

4. **Core memory pinning** - during the session, if the user is working on a specific task that requires persistent context (debugging a complex API, reviewing a long document, planning a project), Claude can PIN critical information to a `PINNED.md` file at the vault root:

• Write task-specific facts, schemas, or reference data to `PINNED.md`
• This file is loaded at L0 alongside SOUL.md and CRITICAL_FACTS.md
• When the task is done, clear `PINNED.md`
• This prevents critical session context from being lost during long conversations or context compaction
• Claude should proactively suggest pinning when it detects the user is deep in a complex task

If identity files (SOUL.md, CRITICAL_FACTS.md) don't exist, offer to create them by asking 5-7 quick questions about the user's role, values, and communication preferences.

If `index.md` doesn't exist, offer to run `/obsidian-init` to generate it.

---

**AI-first rule:** Every note created or updated by this command MUST follow `references/ai-first-rules.md` - `## For future agent` preamble, rich frontmatter (`type`, `date`, `tags`, `ai-first: true`, plus type-specific fields), recency markers per external claim, mandatory `[[wikilinks]]` for every person/project/concept referenced, sources preserved verbatim with URLs inline, and confidence levels where applicable. If that path does not resolve from your working directory, search upward for it; if you still cannot read it, say so before writing rather than producing a note that silently skips the rule. The vault is for future agent retrieval - not human reading.

**Anti-fabrication:** Search exhaustively before claiming any note, person, or file is absent - false absence is the most common failure mode - and never invent facts, entities, or dates (mark unknowns as `TBD`). See the anti-fabrication and search-completeness hard rules in `references/ai-first-rules.md`.